Potential applications of three-way multidimensional scaling and related techniques to integrate knowledge from multiple experts

Theknowledge transfer problem in artificial intelligence consists of finding effective ways to elicit information from a human expert and represent it in a form suitable for use by an expert system. One approach to formalizing and guiding this knowledge transfer process for certain types of expert systems is to use psychometric scaling methods to analyze data on how the human expert compares or groups solutions. For example, Butler and Corter [1] obtained judgments of thesubstitutability of solutions from an expert, then analyzed the resulting data via techniques for fitting trees and extended trees [2]. The expert's interpretation of certain aspects of the solutions were directly encoded as production rules, allowing rapid prototyping. In this paper we consider the problem of combining information from multiple experts. We propose the use of three-way or individual differences multidimensional scaling, tree-fitting, and unfolding models to analyze two types of data obtainable from the multiple experts: judgments of the substitutability of pairs of solutions, and judgments of the appropriateness of specific solutions to specific problems. An application is described in which substitutability data were obtained from three experts and analyzed using the SINDSCAL program [3] for three-way multidimensional scaling [4].     

基于软测量的化工精馏过程推断控制策略

针对化工精馏过程产品成分无法在线检测及其用温度间接控制产品成分的常规控制策略存在着控制精度低的问题,提出基于软测量的精馏过程成分非线性串级推断控制策略。该控制策略首先提出核岭回归的实时软测量方法,即利用满足Mercer条件的核函数改进线性岭回归算法,实现精馏过程产品成分的在线检测;然后在此基础上,提出一种新的非线性串级推断控制策略,即副环采用常规的温度间接控制,主环采用基于核岭回归软测量的推断控制策略。通过Mejedell等建立的精馏塔动态模型分别对单端和双端成分非线性串级推断控制策略性能进行分析,仿真结果表明,与传统控制方案比较,新控制策略的控制质量有了较大提高,控制结构简单,易于实施。     

多重多维模糊推理算法的连续性和逼近性

针对多重、多维模糊推理情形,细致地研究了几类模糊推理算法是否满足连续性和逼近性,并进一步讨论了这几类算法对逼近误差的传播性能。把模糊推理算法看成是一个模糊集合到另一个模糊集合的映射,选用海明距离作为两模糊集的距离度量方法,证明了在模糊假言推理和模糊拒取式推理情形,几类多重多维模糊算法都拥有连续性。当多重多维模糊算法满足还原性时就具有逼近性;该模糊算法都不会放大逼近误差。结果对构建模糊控制系统和模糊专家系统时选用和分析模糊推理算法有一定的指导作用。     

基于码本模型和多特征的早期烟雾检测

提出一种基于码本模型和多特征的早期烟雾检测算法。首先,利用码本模型进行前景提取;然后,结合烟雾的颜色模型和形状特征模型检测前景中的疑似烟雾区域;最后,利用烟雾的动态特性,有效地降低误检率,提高算法的鲁棒性。通过ROC曲线对比,实验结果表明,该算法具有良好的烟雾检测能力。同时,能够满足实时性要求,具有较高的实用价值。     

核空间散度阈值法

提出了一种核空间散度阈值分割方法。首先定义了一种参数型Bregman散度;其次提出了参数型Bregman散度阈值法,并将现有交叉熵阈值法和Otsu阈值法统一起来;再次基于参数型Bregman散度构造了核空间中一种新的不对称核函数,将图像灰度级从欧氏空间变换到再生核空间,获得了一种核空间的散度阈值分割法;最后研究了该分割法中核函数参数选取方法。实验结果表明,基于核空间的散度分割法具有一定的普适性,并能改善现有交叉熵阈值法和Otsu阈值法的分割性能,同时也可将这两种经典阈值分割法看做核空间散度阈值法的特殊情形。     

一种动态场景下运动对象分割新算法

视频运动对象分割是计算机视觉和视频处理的基本问题。在摄像机存在全局运动的动态场景下,准确分割运动对象依然是难点和热点问题。本文提出一种基于全局运动补偿和核密度检测的动态场景下视频运动对象分割算法。首先,提出匹配加权的全局运动估计补偿算法,消除动态场景下背景运动对运动对象分割的影响;其次,采用非参数核密度估计方法分别估计各像素属于前景与背景的概率密度,通过比较属于前景和属于背景的概率及形态学处理得到运动对象分割结果。实验结果证明,该方法实现简单,有效地提高了动态场景下运动对象分割的准确性。     

高比例可再生能源渗透下多虚拟电厂多时间尺度协调优化调度

考虑可再生能源出力、电力负荷和电价等一系列不确定性因素,提出了高比例可再生能源渗透下的多虚拟电厂日内两阶段优化调度模型。该模型通过对多虚拟电厂中运行设备小时级时间尺度和分钟级时间尺度的协调优化,达到分级消除系统中随机和扰动因素的影响,实现高比例可再生能源的安全消纳。并且在多虚拟电厂优化调度模型中,借助Markowitz均值-方差理论,提出利润函数的风险刻画,准确描述不确定性的影响。最后,通过算例分析,说明所提出模型可有效降低决策风险和不确定性因素的影响。     

IgGs-Abzymes from the Sera of Patients with Multiple Sclerosis Recognize and Hydrolyze miRNAs

Autoantibodies-abzymes hydrolyzing DNA, myelin basic protein, and oligosaccharides have been revealed in the sera of patients with multiple sclerosis (MS). In MS, specific microRNAs are found in blood and cerebrospinal fluid, which are characterized by increased expression. Autoantibodies, specifically hydrolyzing four different miRNAs, were first detected in the blood of schizophrenia patients. Here, we present the first evidence that 23 IgG antibodies of MS patients effectively recognize and hydrolyze four neuroregulatory miRNAs (miR-137, miR-9-5p, miR-219-2-3p, and miR-219-5p) and four immunoregulatory miRNAs (miR-21-3p, miR-146a-3p, miR-155-5p, and miR-326). Several known criteria were checked to show that the recognition and hydrolysis of miRNAs is an intrinsic property of MS IgGs. The hydrolysis of all miRNAs is mostly site-specific. The major and moderate sites of the hydrolysis of each miRNA for most of the IgG preparations coincided; however, some of them showed other specific sites of splitting. Several individual IgGs hydrolyzed some miRNAs almost nonspecifically at nearly all internucleoside bonds or demonstrated a combination of site-specific and nonspecific splitting. Maximum average relative activity (RA) was observed in the hydrolysis of miR-155-5p for IgGs of patients of two types of MS—clinically isolated syndrome and relapsing-remitting MS—but was also high for patients with primary progressive and secondary progressive MS. Differences between RAs of IgGs of four groups of MS patients and healthy donors were statistically significant (p < 0.015). There was a tendency of decreasing efficiency of hydrolysis of all eight miRNAs during remission compared with the exacerbation of the disease.     

Comparison of Repeated Doses of C-kit-Positive Cardiac Cells versus a Single Equivalent Combined Dose in a Murine Model of Chronic Ischemic Cardiomyopathy

Using a murine model of chronic ischemic cardiomyopathy caused by an old myocardial infarction (MI), we have previously found that three doses of 1 × 10⁶ c-kit positive cardiac cells (CPCs) are more effective than a single dose of 1 × 10⁶ cells. The goal of this study was to determine whether the beneficial effects of three doses of CPCs (1 × 10⁶ cells each) can be fully replicated by a single combined dose of 3 × 10⁶ CPCs. Mice underwent a 60-min coronary occlusion; after 90 days of reperfusion, they received three echo-guided intraventricular infusions at 5-week intervals: (1) vehicle × 3; (2) one combined dose of CPCs (3 × 10⁶) and vehicle × 2; or (3) three doses of CPCs (1 × 10⁶ each). In the combined-dose group, left ventricular ejection fraction (LVEF) improved after the 1st CPC infusion, but not after the 2nd and 3rd (vehicle) infusions. In contrast, in the multiple-dose group, LVEF increased after each CPC infusion; at the final echo, LVEF averaged 35.2 ± 0.6% (p < 0.001 vs. the vehicle group, 27.3 ± 0.2%). At the end of the study, the total cumulative change in EF from pretreatment values was numerically greater in the multiple-dose group (6.6 ± 0.6%) than in the combined-dose group (4.8 ± 0.8%), although the difference was not statistically significant (p = 0.08). Hemodynamic studies showed that several parameters of LV function in the multiple-dose group were numerically greater than in the combined-dose group (p = 0.08 for the difference in LVEF). Compared with vehicle, cardiomyocyte cross-sectional area was reduced only in the multiple-dose group (−32.7%, 182.6 ± 15.1 µm² vs. 271.5 ± 27.2 µm², p < 0.05, in the risk region and −28.5%, 148.5 ± 12.1 µm² vs. 207.6 ± 20.5 µm², p < 0.05, in the noninfarcted region). LV weight/body weight ratio and LV weight/tibia length ratios were significantly reduced in both cell treated groups vs. the vehicle group, indicating the attenuation of LV hypertrophy; however, the lung weight/body weight ratio was significantly reduced only in the multiple-dose group, suggesting decreased pulmonary congestion. Taken together, these results indicate that in mice with chronic ischemic cardiomyopathy, the beneficial effects of three doses of CPCs on LV function and hypertrophy cannot be fully replicated with a single dose, notwithstanding the fact that the total number of cells delivered with one or three doses is the same. Thus, it is the multiplicity of doses, and not the total number of cells, that accounts for the superiority of the repeated-dose paradigm. This study supports the idea that the efficacy of cell therapy in heart failure can be augmented by repeated administrations.